John Day: “IPv6 is still not the solution”

The professor and pioneer Internet researcher, John Day, assures at the i2CAT Foundation in Barcelona that “a large part of the problems of the TCP/IP architecture (multi-homing, mobility, scalability in routing) come from the fact that the IP address identifies an interface on a node instead of the node itself.”

The i2CAT Foundation has begun working with Boston University, the Irish research center TSSG and the American start-up TRIA Network Systems on the development of the first prototype based on the RINA architecture (Recursive InterNetwork Architecture), a general theory of network architecture developed by Internet pioneer John Day. This professor and researcher from Boston University has recently been invited by i2CAT to Barcelona to contribute his vision of the Internet of the Future, a computer network much simpler than the current one and, therefore, cheaper to build and operate, while providing multi-homing, mobility, multicast and quality of service to users.

John Day, during his visit to Barcelona, has stated that the current Internet is still an incomplete demo of what a computer network should be: “it is a very close reflection of the first prototypes of packet switching networks built during the 70s, with the big difference that the hardware has improved noticeably; what's more, some network architectures of that time that did not prosper (such as Cyclades, XNS, DECnet or OSI) had fewer structural limitations than the current architecture based on TCP/IP protocols.”

According to the theorist, much of the problems of the TCP/IP architecture (multi-homing, mobility, scalability in routing) come from the fact that the IP address identifies an interface on a node instead of the node itself. Because IP networks compute routes as a sequence of IP addresses (and therefore as a sequence of interfaces), if a node is connected to the network through more than one interface (multi-homing), the network has no way of knowing that all of those interfaces belong to the same node. Consequently, when one interface of the node fails, the network does not automatically send packets through another interface of the same node, but rather discards it.

“Taking into account that IPv6 is based on the same protocol and continues to identify the interface and that's it, we find ourselves with the same problem,” says Day, who points out that “IPv6 does not provide any solution to the Internet of the Future, it is just another patch on the current Internet, already saturated.”

What the professor proposes with his RINA architecture is a model in which a layer is a distributed application that provides communication services between processes through a certain scope (such as a point-to-point link, a local network, a regional network, a network of networks...). These layers (called DIFs in RINA terminology) are recursive, since they provide services to each other, and as many as the network designer considers appropriate can be used. That is, in contrast to the current model, we would be talking about a single type of layer (configurable depending on where it operates) that can be repeated as many times as necessary to better manage network resources. The current model, based on a fixed number of functional layers (each layer performs a different function) is more complex and inefficient due to the multitude of protocols it needs, inflexible and, in addition, the implementations do not follow the current theoretical model (protocols that are between two layers, repeated functions in different layers, interactions between layers that should not happen).

The i2CAT Foundation will contribute to making this theory a reality, developing together with the aforementioned international entities the first prototype of this revolutionary architecture. This project aims to validate and improve the specifications of the protocols that implement RINA, also resulting in two independently developed but interoperable prototypes. Another important objective of the project is to demonstrate that a complex transition process is not necessary to adopt the RINA architecture, but that in the future RINA can begin to be deployed coexisting with TCP/IP networks. Its results will lay the foundation for better implementations to continue developing and experimenting with the properties and benefits that RINA entails.

The i2CAT Foundation is a research and innovation center founded in 2003. Its main objective is to achieve an accessible and open Internet of the future. To this end, the Foundation makes Internet research and innovation available to society, through collaboration between the public administration, companies and research groups from universities and the educational world for the development of innovative projects in the field of new technologies of the knowledge society.